It seems to work better now… I think the problem was perhaps related to Solver=Pardiso.
DId you get this erro too?? Can you guide me in a way to solve this?
Yes, I mentioned it here: .inp file details - #18 by FEAnalyst
Correct. I was asking about any tip in how to solve it.
Indeed, I have always done these type of simulations as the BC how I propossed them here, and i did not have convergence problems… so I am a bit blocked on how to solve this…
I saw a tutorial where you applied a foce to a distance of the Ref. Node in order to generate a moment and said that it had a better convergence… however not sure how to do it?
Perhaps it is not the place to ask, in which case tell me, but if you could please guide me or give me any example similar to what i work with which I can consult. I work with gear simulations, so always apply torue in the pinion and a rotation to the gear. In this specific case I did not have a gear and had a rack so did not apply the rotation.
Thank you very much!
Thank you very much
This thread on the CalculiX forum can be helpful: Zero reaction moment when reading from ROT NODE in 2D - CalculiX (official versions are on www.calculix.de, the official GitHub repository is at https://github.com/Dhondtguido/CalculiX).
Basically, prescribed displacement/rotation is always better for convergence than force/moment and you should avoid rigid body motions in static analyses. Of course, there are workarounds like running a dynamic analysis or adding some soft springs to stabilize the model but it’s sometimes better to just assume the worst case and fix one gear while rotating the other.
1 Like